Abstract Abnormal RNA splicing is an underrecognized driver of pathogenicity in germline TP53 — the cause of Li–Fraumeni syndrome (LFS). We re-evaluated exonic single-nucleotide variants (SNVs) that yield missense or synonymous changes for spliceogenic effects by integrating SpliceAI prediction, in-vitro minigene assays, and analysis of tumor RNA-seq from TCGA, and assessed genotype-phenotype correlations using clinical data from multiple databases and national registries. We identified 58 spliceogenic exonic SNVs (SE-SNVs) across the TP53 gene (40 missense, 18 synonymous). Experimental validation confirmed aberrant splicing for 15 out of 17 tested variants, most often through cryptic splice-site activation that introduced frameshifts and premature termination. Clinically, carriers of SE-SNVs previously considered as mild or of low-pathogenicity by protein-based assays showed earlier onset and LFS-signature cancers, indicating that splicing disruption can override amino-acid effects. The recurrent c.375 G > A (p.(Thr125 = )) showed heterogeneous effect: with both childhood/adolescent and adult onset, consistent with partial, variable retention of canonical splicing. These data reveal a substantial burden of spliceogenic pathogenicity in TP53 and strong support integrating splicing prediction, functional validation, and transcript-level evidence into variant interpretation and risk stratification in LFS.
Schönegger et al. (2026) studied this question.